Excavation of a Romano-British Settlement on the A149 Snettisham Bypass, 1989 0905594312, 9780905594316

Citation preview

EAST ANGLIAN ARCHAEOLOGY

Excavation of a Romano-British Settlement on the A149 Snettisham Bypass, 1989

by Myk Flitcroft

with specialist contributions from S. Anderson, T. Ashwin, S. Cottam, J.A. Davies, B. Dickinson, Tony Gregory, David Gurney, J. Hillam, D. Mackreth, G. McDonnell, C. Mortimer, P. Murphy, J. Price and D. Williams and illustrations by Steven Ashley, Dave Fox and Piers Wallace

East Anglian Archaeology Report No.93, 2001

Archaeology and Environment Division Norfolk Museums and Archaeology Service

EAST ANGLIAN ARCHAEOLOGY REPORT N0.93

Published by Archaeology and Environment Division Norfolk Museums and Archaeology Service Union House Gressenhall De re ham Norfolk NR20 4DR in conjunction with The Scole Archaeological Committee

Editor: Peter Wade-Martins Managing Editor: Jenny Glazebrook

Scale Editorial Sub-committee: Brian Ayers, Archaeology and Environment Officer, Norfolk Museums and Archaeology Service David Buckley, County Archaeologist, Essex Planning Department Keith Wade, Archaeological Service Manager, Suffolk County Council Peter Wade-Martins Stanley West Set in Times Roman by Joan Daniells and Jenny Glazebrook using Core! Ventura™ Printed by Witley Press Ltd., Hunstanton, Norfolk ©ARCHAEOLOGY AND ENVIRONMENT DIVISION, NORFOLK MUSEUMS AND ARCHAEOLOGY SERVICE ISBN 0 905594 31 2 For detail s of East Anglian Archaeology, see last page This volume is published with the aid of a grant from English Heritage

Cover photograph: Aerial view of the excavation in progress, 10 October 1989 Photo by D.A. Edwards (ref: TF6733/W)

Contents Contents List of Plates List of Figures List of Tables List of Contributors Acknowledgements Summary/Resume/Zusammenfassung

Well2 Enclosure 5 Enclosure 6 House 3 Corn Drier 144

V VI

vi vii vii V.

VII

viii

Trackway Ditches 274 and 297 Enclosure 7 Grave 549 Other Features

Chapter 1. Introduction I. Il.

Ill.

IV. V.

Introduction The Geographical Setting Geology Topography The Archaeological Background The Region The Ingol valley The Snettisham Bypass Project Format of the Report

I 1 I 1 1 1 2 5 5

VI.

II.

m.

Introduction Coverage The Field Survey Res ults Field walking Soi l Surveys, by David Gurney Discussion

Phase 5 (third century AD) Description Well! Other Features

Chapter 5. Watching Brief Results I. II. Ill.

Chapter 2. 1983- 1984 Fieldwork I.

Phase 4 (late second century AD) Description

6 6 6 6 6 7

Introduction Watching Brief Methodology Watching Brief Results Sites 1487, 1544 Site 25750 Site 25751 Site 18236 Site 1515 Description Finds Summary

Site 20199

Chapter 3. The 1989 Excavation: Site 1555 I. II. Ill.

Introduction Site 1555 Excavation Methodology Phas ing of Contexts

II.

Non-Roman Findings Prehistoric Finds Post-Roman Features Romano-British Activity: Phase 1 (mid first century AD) Description Enclosure 1 House 1 Enclosure 2

Discussion : Finds and Environmental Evidence Other Features Industrial Waste Pit 730 Hearth 155

Ill.

Romano-British Activity : Phase 2 (late first century to early second century AD) Description Overall Site Layout House 2 Enclosure 2 Well!

IV.

Phase 3 (early-mid second century AD) Description Trackway Ditches 245 and 285 Trackway Ditches 57 I and 575 Enc losure 4

47 47 47 47 47 47 47 49 49 49 50

Chapter 6. Finds Summaries

9 9 12

I. II.

Chapter 4. Excavated Features I.

32 37 37 37 37 42 42 42 42 42 42 44 44 44 46

Ill. IV.

15 15 15

V. VI. VII. VIII. IX.

15 16 16 16 20 20 21 21 22 22 22 22 23 28 29 31 31 31 31 32

X. XI.

Iron Age Coins, by Tony Gregory Roman and Later Coins, by John Davies Roman Issues Post-Roman Coinage Copper Alloy Brooches, by D.F. Mackreth Metalworking Debris Slag, by Gerry McDonnell Moulds and Crucibles, by C. Mortimer Glass, by J. Price and S. Cottam Samian, by B. Dickinson Mortaria, by David Gurney Amphorae, by D. Williams Coarse Pottery, by M. Flitcroft Introduction Analysis Methodology Fabrics Forms Residuality and Redepositon Phase Ceramic Summaries Discussion Brick and Ti le, by M. Flitcroft Daub/B urnt Clay, by M. Flitcroft Loomweights

51 51 51 52 52 55

55 56 58 59 59 61 63 63 63 63 64 65 66 70 74 74 75

Chapter 7. Environmental Evidence I.

11.

V

Human Bone, by S. Anderson Inhumation 548 Juvenile Mandible Discussion Animal Bone, by T. Ashwin

76 76 76 76 76

Ill.

Plant Macrofossils, by. P. Murphy Results

77 77 77 77 77 77 77 77 77 77 78

The Wells Corn Dri er 144 The Houses

Conclusions Dendrochronology, by J. Hillam Results

IV.

Well! Well2

Discussion

Chapter 8. Discussion I.

The Local Setting: The 1989 excavations in relation to the Ingol valley The Regional Setting

11.

79 80

Appendix 1. Structure of the Archive Report

82

Bibliography Index , by Sue Yaughan

84 86

List of Plates Pl.I PI. II PI. Ill PI. IV PI. V

Topsoil removal in progress on Site 1555 8 Cropmark coverage on Site 1555, to east of excavation area 10 Aerial photograph of Site 1555 during 11 excavation House 1 before excavation 18 House 2 and enclosure after excavation 23

PI. PI. PI. PI. PI. PI.

VI VII VIII IX X XI

Well 2, showing details of lining Corn Drier 144 Inhumation 548 Well 1, showing details of lining Well 1, dismantling of timber lining Iceni gold quarter stater

36 40 42 45 45 51

List of Figures Fig. 1 Fig. 2 Fig. 3 Fig. 4

Fig. 5 Fig. 6 Fig. 7 Fig. 8 Fig. 9 F ig. 10 Fig. 11 Fig. 12 Fig. 13 Fig. 14 Fig. 15 Fig. Fig. Fig. Fig. Fig.

16 17 18 19 20

Fig. 21

Location of Snettisham Location of 1989 excavations and other known Romano-British sites Romano-British settlement cropmarks in the Ingol valley Distribution plot, showing main findings of 1984 fieldwalking and geochemical concentrations Site plan, showing all excavated features Site plan, Phases 1 and 2 Prehistoric finds Sequence of Phase 1 structures around House 1 Plan of House 1, showing structural features House 1, sections across structural features House 1, sections across non-structural features House 2 and associated enclosure House 2, sections across ring ditch House 2, sections across enclosure ditch Distribution of finds around House 2 ring ditch and enclosure Pl an of Well 1 Isometric reconstruction of Well 1 lining Sections across Well s 1 and 2 Site plan, Phase 3 Plan of Enclosure 4 and associated features Sections across Enclosure 4 and associated features

2 3

Fig. 22 Fig. 23 Fig. 24

4

Fig. 25 Fig. 26 Fig. 27 Fig. 28 Fig. 29 Fig. 30 Fig. 31 Fig. 32

7 13 14 15 17

Fig. 33 Fig. 34 Fig. 35

18 19

Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig.

20 24 25 26 27 29 29 30 33

36 37 38 39 40 41 42 43 44

Fig. 45 34 34

vi

Plan of Well 2 Isometric reconstruction of Well 2 lining Plan of Enclosures 5 and 6, and associated features Sections across Enclosures 5 and 6, and associated features Plan of corn drier 144 Sections across corn drier 144 Site plan, Phases 4 and 5 Site 1515 , site plan Brooches from House 2 enclosure ditch Brooch assemblage, Site 1515 Copper alloy mould fragments from pit 730 Glass vessels, Site 1555 Illustrated samian Mortaria assemblage, Sites 1555 and 1515 Phase 1, pottery assemblage House 2, pottery assembl age Phase 2, pottery assemblage Phase 3, pottery assemblage Phase 3, pottery assemblage Phase 4, pottery assemblage Well 1, pottery assemblage Site 1515, pottery assemblage Pottery fabric proportions by phase and total Key to conventions used in plans and sections

35 36 38 39 40 41 43 48 53 54 57 58 60 62 66 67 68 69 71 71 72 73 74 81

List of Tables Table I Table 2

Chronological breakdown of the coins from Snettisham Bypass Breakdown of first and second century coin denominations

Table 3 Table 4

51

XRF analysis of moulds and crucibles Coarse pottery fabrics; summary of sherd mass for all phases

57 65

52

List of Contributors Sue Anderson, BA, MPhil, MIFA

Jennifer Hillam, BSc, FSA, MIFA

Human Bone Specialist

Dendrochronology Laboratory, Department of Archaeology and Prehistory, University of Sheffield

Trevor Ashwin, BA Senior Project Manager, Norfolk Archaeological Unit

Gerry McDonnell, BTech, PhD

Sally Cottam

Department of Archaeological Sciences, University of Bradford

formerly Department of Archaeology, University of Durham

D.F. Mackreth, BA, FSA Freelance Consultant

John A. Davies, BA, PhD

C. Mortimer, BTech, DPhil

Keeper of Archaeology, Norwich Castle Museum

Freelance Consultant

Brenda Dickinson

Peter Murphy, BSc, MPhil

Freelance Consultant

Centre of East Anglian Studies, University of East Anglia

Myk Flitcroft, BA, MSc, MIFA formerly Senior Project Manager, Norfolk Archaeological Unit

Jennifer Price, PhD, FSA Department of Archaeology, University of Durham

The late Tony Gregory, MA, AMA, MIFA

D.F. Williams, PhD, FSA English Heritage Ceramic and Lithic Petrology Project, Department of Archaeology, University of Southampton

David Gurney, BA, MIFA Principal Landscape Archaeologist, Norfolk Landscape Archaeology

Acknowledgements for getting the report itself out of my head and onto paper, and for correcting my grammar once it had been written. I would like to thank David Gurney for his overview and advice on the fieldwork and early analysis stages of the project, and for discussion and his comments on drafts of the report text; I am also grateful for the positive critici sm provided by Jude Plouviez and Liz Shepherd on previous drafts. The Universities Superannuation Scheme, landowners on main excavation site, are to be thanked for their donation of the finds to Norfolk Museums Service; their tenant on the excavation site, Mr Keene, for assisting with early access to main excavation site and providing help whenever it was asked. The quantities of metalwork and coins from the topsoil were recovered due to the assistance of many local amateurs who worked along side the excavation team: I would like to thank all the metal-detector users and local amateurs who volunteered during the course of the excavation , but particul arly John Bocking, who was an ever-present help.

The greatest debt is to the excavation and post-excavation team: Jonathan Erskine (Site Supervisor), Phi! Copleston (Finds Supervisor), Melanie Gauden, Dave Hodgekinson, Mike Hum, Michael Ings, Jamie Patrick, Jens Samuels, Simon Savage, and Lisa Wastling (Excavators); Steven Ashley, Dave Fox, and Piers Wallace (Illustrators); and all the speciali st contributors to this report. The fieldwork stage of the excavation was funded by Norfolk Cou:1ty Council, while the post-excavation funds were provided by English Heritage: thanks are due to Phi lip Walker (EH) for his championing of the analysis programme and to Charles Auger (NCC) for arranging early access to the site. I am grateful to all my colleagues in the NAU for their advice on aspects of this report, but particularly to Trevor Ashwin and Liz Shepherd for their valuable advice on structure and format of the report, and to A lice Lyon s for her thoughts on the pottery. Jez Reeve and Brian Ayers were invaluable for their suggestions on the tone of the report,

VII

Summary Part of a Romano-British settlement to the west of the village of Snettisham was excavated in 1989 prior to the construction of the A 149 bypass. Although the excavation area was only a small part of the extensive Romano-British settlement in the Ingol valley, valuable information concerning the nature of the activity was recovered. The excavation revealed that truncated remains of an extensive settlement dating from the mid-first to the late second century survived beneath the modem ploughsoil. The Romano-British settlement was based on a mixed economy of farming and low intensity industry and demonstrated the survival of traditional techniques of house construction, and the continued importance of

handmade pottery well into the Roman period. The excavation produced a useful pottery assemblage which complements other groups from the Saxon Shore Fort at Brancaster (Hinchliffe and Sparey-Green 1985), and Fenland sites. The decline in valley floor activity occurred at approximately the same time as the construction of a villa complex on higher ground to the east, and it is possible that environmental changes in the Fenland region during the third century resulted in the shift of settlement to the east. This eastern focus was also used in the Saxon and medieval periods, and no evidence for intensive post-Roman use of the excavation area was found.

Resume provenant du Saxon Shore Fort a Brancaster, et de plusieurs sites des Fens. L'activite se reduisit dans la vallee environ a la meme epoque que la construction d'une villa sur une hauteur situee a l'est, et il est possible que des modifications de l'environnement dans la region des Fens au cours du troisieme siecle entrainerent un deplacement des habitations vers I' est. Ce changement persista pendant les periodes saxonne et medievale, et nous ne possedons aucune preuve d'une utilisation intensive de la zone fouillee dans la periode post-romaine. (Traduction: Didier Don)

Les terres de labour de Snettisham recelent encore en leur sous-sol quelques traces d'une importante implantation qui s'etendait du milieu du premier sieclejusqu'a la fin du deuxieme siecle. Fondee sur une economie mixte, reposant a la fois sur I' agriculture et sur une industrie de faib le envergure, cette implantation montre que des techniques traditionnelles de construction de maisons subsistait encore a cette epoque et que la poterie faite a la main conservait de l'importance alors que la periode romaine etait deja bien entamee. Les fouilles ont permis de mettre a jour un ensemble de poteries d'une grande utilite qui viennent completer d 'autres ensembles

Zusammenfassung angelsachsischen Kiistenfort von Brancaster und verschiedener anderer Statten im Fenland erganzen. Die Aktivitaten im Talgrund lieBen zur etwa gleichen Zeit nach , da weiter ostlich eine hoher gelegene romische Villa erbaut wurde. Es ist durchaus moglich, dass okologische Veranderungen im Fenland im 3. Jahrhundert zu einer Siedlungsverlagerung nach Osten hin fiihrten . Die Ausrichtung nach Osten setzte sich in der Zeit der Angelsachsen und im Mittelalter weiter fort. Die Grabungsstatte ergab keine Hinweise auf eine intensive Nutzung in nachromischer Zeit. (Obersetzung: Gerlinde Krug)

Teilwei se erhaltene ausgedehnte Siedlungsreste aus der Zeit Mitte des 1. Jahrhunderts bis Ende des 2. Jahrhunderts haben unter der heutigen Ackerkrume bei Snettisham iiberlebt. Die romisch-britische Sied lung, die auf einer Mischung aus Landwirtschaft und leichter Industrietatigkeit aufbaute, demonstriert das Fortbestehen traditioneller Techniken im Hauserbau und die bis weit in die Romerzeit hinein anhaltende Bedeutung handgefertigter Topferwaren. Die Grabung fOrderte eine Reihe brauchbarer Keramikgegenstande zutage, die Funde aus dem

VIII

Chapter 1. Introduction I. Introduction The construction of the 5.4km long DersinghamIngoldisthorpe-Snettisham Bypass in 1989-1990 followed many years of discussion over its exact course. It was apparent, however, that the road was likely to affect areas of potential archaeological interest and in order to assess the ex tent of this disturb a nce the Norfolk Archaeological Unit conducted a fieldwalking and geochemical survey of the proposed roadlines in 1983. This programme identified the section of the road due east of Ingoldisthorpe village as the primary area of interest, where the bypass would cross the apparent site of a Romano-British settlement. The archaeological response to the road building project was two-fold. A Watching Brief was to be kept on soil disturbance along the whole of 5.4km of the bypass line from TF 683293 to TF 683347; in addition more intensive excavation was proposed to study the 800m length spanning the boundary be~ween Ingoldi sthorpe and Snettisham pari shes between a track known as 'The Drift' at TF 67888 32455 and the river Ingol at TF 67585 33220. Negotiations between the Unit, the developer (Norfolk County Council), and English Heritage resulted in the Council agreeing to fund the 1989 fieldwork while English Heritage met the post-excavation costs.

11. The Geographical Setting Geology The geology of the area comprises part of the Cretaceous Greensand series (white Sandringham Sands, overlain to the east by the loams and clays of the Snettisham Clays, and the ferruginous Carstone) (Casey and Gallois 1973). Trial pits dug on the main excavation site showed that the white Sandringham Sands lay beneath an iron-stained upper zone of sands, covered by between 0.4m and 0.6m of silty grey sand topsoil. Environmentally the area lies on the edge of a northern extension to the silt skirtlands of the Fens between the Carstone and chalk uplands to the east and the mud flats of the Wash to the west.

Topography The main topographical feature which the bypass crosses is the broad and shallow valley of the river Ingol which cuts down from the uplands to empty into the Wash. The modern river has been canalised west of Snettisham, and its original course is unknown. Study of Ordnance Survey ma ps suggests that the parish boundary between Snettisham and Ingoldisthorpe may preserve an earlier course but no trace of a relic stream channel was encountered in the main excavation area and it is possible that the stream ran to the south of 'The Drift' . The old A 149 follows the 1Om contour around the base of the escarpment to link the modern villages ; the bypass cuts straight across the valley floor, which dips to a little below 5m OD. This route brings it to the edge of the more

marshy zone, and this is particularly noticeable at the southern end of the route, where the road cuts through Dersingham Bog, a Site of Special Scientific Interest (SSSI). The main excavation area lies on a slight south-facing rise on the northern side of the valley floor. The ground rises from around 5.5m to 6.8m OD in the southern 240m of the site, but is generally flat for the remainder. The modern water table was found to be about 1.5m below the field surface, but local information states this is a result of land drainage over the last forty years, and that prior to this the land was very boggy and subject to intermittent upwelling of springs.

Ill. The Archaeological Background (Figures 1-3)

The Region The nature of Romano-British settlement in North-west Norfolk has been discussed by Tony Gregory (Gregory 1982), and subsequent work has not radically altered his description of the area (Figure 1). The area was served by communication routes running north south and east-west. The Roman road, Peddars Way, ran to the east of the main settlement concentrations towards a probable ferry terminal at Holme next the Sea; the prehistoric communication corridor of the Icknield Way followed a similar route further west, rather nearer to the excavation area, and this route continued in use throughout the Roman period (as the location of the settlements along its length indicates). The east-west road known as the Fen Causeway crossed the fens from Durobrivae entering Norfolk at Denver, where it crossed a second north-south route, Akeman Street. This last road's route further north cannot be traced, but it seems to have followed the fen edge at least as far as Shouldham and the Ri ver Nar. Settlement in this area was concentrated along the west side of the chalk uplands, between the chalk and the coastal marshes or fens. Within thi s band valley-side or spring- line locations were preferred, where a variety of soil-types (and hence environments) could be exploited. The chalk uplands to the east appear to have been less densely settled. This pattern led to a chain of romanised buildings along the chalk escarpment, extending from south of Snettisham to Gayton, 15km away (the seven 'Icknield Way villas' referred to by Tony Gregory (Gregory 1982, 360-366) to which the Snettisham Park Farm villa might be added (Site 1514 on Figure 2). Apart from this last site (with the Ingol settlement to the west and some form of occupation at the head of the Ingol valley at Shernbourne), none of these settlements have provided great evidence for satellite settlements and were interpreted as small estates completely based on the villas themselves. The apparent lack of luxury on all these sites in West Norfolk may be a result of the marginal location of the area in relation to the social centres of the civitas capitals and later to the economic network of 'small town s' that

I : ,\

Heachame \\ Snettisha ·~· \ \•Fr i ng

::·.':. I I \\

.·:··

.·:'.' · · ·.

\

~

•West Newton • '-. \ \ Toftrees

"-

\"" -.(... •

Sa ham Toney

,~

\'!,.. e.

\

(

~J £. •\ Brettenham

'•~

lOOk m

A

f===9

0

50miles 0 0

~

20km lOmile s

Figure 1 Location of Snettisham grew up across Roman Britain fro m the second century onwards. The area lies approximately 70km from the civitas capital of Venta lcenorum (Caistor St Edmund), and around 30km from the closest of the smaller towns (at Toftrees, Saham Toney, Billingford , and Brettenham- on Figure 1); this latter di stance is more than twice that found in other areas (M ili eu 1990). The cropmark ev idence surroundin g the Fen Causeway around Denver does suggests a settlement of suitable size to act as a market centre, but there is no evidence to confirm thi s function.

there is not the density of finds or findspots to suggest that thi s marks a continuation of the settle ment zone itself. The area of occupation appears to have been limited to the north by the slope of Ken Hill: little Romano-Briti sh material had been prev iously found in thi s area and this impression was not altered during the 1989 Watching Brief on either Site 1544 or Site 1487 . The southe rn boundary of settleme nt is unclear: Sites 201 99 and 11 829 indicate that Romano-British cultivation continued to the south of the area excavated, but exactly how far south is not known as no cropmarks or surface fi nds are recorded in the county SMR for the area south of thi s. The area was unavail able during the 1983 initi al fieldwalking survey, but no new findin gs were made during the 1989- 1990 Watching Brief. Material recovered fro m the valley fl oor sites con sisted, in the main , of a range of Romano-B ritish ceramics, bri ck and til e, co in s and brooches; find s of iron s lag we re also reported from many of the locati ons, principally those to the north and east of Site 1555. Evaluation of fields to the north of Common Road, Snetti sham in 199 1 (Site 28450 on Figure 2) demonstrated a similar ra nge of material survived in this north-eastern zone, but also identifi ed metalled road surfaces and two early Romano-Bri tish kilns (Flitcroft 199 1). These finds were made close to the area whi ch had pre vi ously produced the ' Je wellers Hoard ' (SMR Site 15 17), a coll ection o f gold and sil ver unearth ed durin g buildin g work, a nd interpre ted as a hoard of waste obj ects collected fo r re-use (Potter 1986) .

The Ingol Valley Study of the county Sites and Monuments Record (SMR) shows that the main excavati on area (Site 1555) is situated towards the southern end of a widespread area ofRomanoBriti sh f ind-spots. Sites produc in g Roma no-Briti sh material are shown on Figure 2; cropmark coverage for the immedi ate area of the 1989 excavation is detailed on Figure 3. It ap pears that much of the northern part of the Ingol valley was under cultivation, with small rectangular fields in terspersed with linear tracks (SMR Sites 151 5, 18236, 1543, 1554, 1556, 20 199, 11 829, 28450) and an ex tensive occupation zone to the immedi ate east of the main excavation area (S ite 1555). Surface find s reported to the county SMR suggest a Romano-British date fo r these cropmark compl exes, but also extend the area of Romano-Briti sh acti vity to the west into the lower-lying marshy area (e. g. Sites 1524, 1525, 24583). Although

2

34 1503 25779

25797 1525

33

1524

~ N

t

31

SNETTISHAM BYPASS 1989 0

500

1000 m

========~=======

66

67

Figure 2 Location of 1989 excavations and other known Romano-British sites. Scale 1:25,000

3

34 250

0

500m 28450

~ N

t

W-1

---33

SNETTISHAM BYPASS 1989 1555 ING Settlement Cropmarks

Q)

0\

00

0\

.s ::0 C

a

f

e

N

g

(>

.

-

.1

c l_

~--

- ·--

_

I

I

-~ - - -

__ _ _

_ _J

~ I - . -= j:--

I -~ - -

I

--= j--

0 35 150

0 35 200

/

/

\ /

\

/

\

-

I

Phase 3a

~P hase 3b

I

---

,,

'

I

-,-:::; __

I

_

-----=- i-- - 03 6 550

036 600

I

_ 1_ 1-

- - -i =- -- o- -

-

_ I_ 1- -

I

I

__ ---:-j= -

_ ---:j =036 500

036 450 0

Figure 19 Site pla n, Phase 30 Sca le I :500

33

10

20m

I

- _ --:--j=-

I

_-: j=-

0- . -

I

--=-i-- - -

I

.=c_

I

. --=i:--- .

I

0

_

--=-T 0 36 400

I

---=-c .

SNETTISHAM BYPASS 1989

Site 1555

0351450

050 1450

- i-1

- i-

I I

I I

I

I I I

I

I

I

0351 440 - i- I

~:!_40

_J

I L

543

I

I

I I

I 589I 1 I

15ss

I

I I I

I

I

I

I

I

Enclosure 4

I I I

1588 I I I I

i

I

0351430

I 0501430 - i-

I

- i-

I

I I I

I

I

I I I I I I I I

i

I

I

I

I

I I

\

I 605

'"·,j

I

.

I I

I I

i I I

I I

i

I

\)533

I

I

I

I

i

I

050 1420

03~_,:!_20

- i-

I

5

0

10 m

Figure 20 Plan of Enclosure 4 and associated features. Scale 1: 125

Site 1555

SNETTISHAM BYPASS 1989

s

N

E

W

[>



v;-~ ··

.. ··

\:

·_:::··· ·

I

';

(

~

.I

.....-

I ./

,

33

I

.· :/.40~ - - -~ .•'

I I

,··

I I

:

16

.·1 .I 1

........ ··

I

• 28

I

············

.-~---·

:~ :: ~

I

1 1

... ···

35__..- -- -:;'(· ..

I

... ···

•

I

39

I I

.ts

~

• 22

N

t

_11} 3 6751

-..r

Recorde d features Road lin e

I I I I I I I

Crop m arks plotted

0

50 m

/ /

/

/ / /

I I I _I

/

/

/ - ----

/

I

/

I I

I I

/

Figure 29 Site 1515, site plan. Scale 1:1000

48

/

/ /

18, 27, 36, 37, Curie 11; Central Gaulish Forms 18/31R, 31, 33, 37, 38, 81 and East Gaulish (La Madeleine factory) Form 27 represented. A full descriptive catalogue is included in the Archive Report (Section 4.7).

Site 1515 (Figures 2 and 29) The only place in which extens ive areas of archaeology remained was immediately north of the main excavation (Site 1555), on the opposite bank of the modem course of the In gol. This area also had cropmark coverage and had been previously fieldwalked, and the excavated features correlated reasonably well with these. Figure 29 records the location of all the features seen on this site, with cropmarks plotted as dotted lines. The features here had to be recorded quickly, and a salvage record was all that could be made.

Mortaria by David Gurney Parts of five mortaria were recovered from the fills of trackway ditches 16 and 34 on Site 1515. Four of the vessels are likely local products from unidentified kilns, having a cream/yellow fabric containing very sparse quartz and reddish-brown inclusions and flint and/or quartz trituration grits (Archive Report Nos 34,. 35, 36, 38). The fifth vessel (Figure 35, No. 33) ts more interesting, being of a different fabric and also stamped. The stamp is clear enough, but not entirely straightforward to interpret. It is likely that the name Crescens is intended. One potter's stamp survives and two others are recorded, both from Baylham, near Ipswich. There is a good possibility that he worked at Baylham, and the rim-forms leave no doubt that he was active in the period AD60-90 (K. Hartley pers. comm.). The date range of these five vessels is uncertain, but appears to cover the period from the mid first century (No. 33, above) to the second century. The fabrics represented are very similar to those from Site 1555, and it can be assumed that mortaria were made at the same ?East Anglian workshops. A full descriptive catalogue is included in the Archive Report (Section 4.8).

Description (Figure 29) Two ditches (16 and 34) formed the sides of a major trackway running between fields and enclosures, with two others (36 and 37) marking a subsidiary lane running off thi s track. The oth~r ditches recorded (33, 38, 40) could not be related to the cropmark pattern, and appear to be a later reorganisation. The area south of the major track (16/34) may have been used for industrial processes as five small pits (14, 15, 22, 28, 39) were recorded on the roadline in this area, and their fills produced slag and pottery wasters. The in fill of the southern major track ditch (16) also suggested pottery manufacture in the southern part of this field, with large numbers of partially complete vesse ls and wasters be ing recov ered. The pottery assemblage from these ditches was comparable with that from Site 1555, although wasters were only found on Site 1515. Samian and mortaria from the ditch suggest this production was centred in the mid second century, which again compares well with the results from Site 1555.

Amphora by D. Willi ams Sherds from at least four vessels were recovered (none illustrated). Examples of Dressel 20, Gauloise 4 and a ?Gallic type were identified.

Finds Summary: Site 1515 Apart from Site 1555, this part of the bypass line was ~he only area to produce significant quantities of mate.nal. Most of it was recovered during the salvage excavattons and provides an interesting complement to the artefacts from the main site.

Pottery by M . Flitcroft A total of 12.4kg of coarse pottery was recovered during salvage excavations on Site 1515, almost all of this identified as Romano-British material. As the site was excavated under adverse conditions and no phasing of the contexts was possible, the pottery has been considered as a site total and not subdivided in any way. The range of fabrics present is very similar to that on Site 1555, with the gritty local reduced ware R4, and the grey wares G2 and G5 being the most common. No Nene Valley wares were identified and this, taken in conjunction with the absence of later vessel form types, may indicate an earlier cessation of activity in this part of the settlement. Forms include types intuitively considered as early Roman , such as campanulate bowls; types similar to the Iron Age style handmade vessels found on Site 1555 are absent. Other vessel types include a range of jars, and bowls and folded beakers in local reduced fabrics similar vessels were found during an evaluation of a kiln site 700m east (Site 28450) within the same settlement zone (Flitcroft 1991). Decoration consists of external burnishing, scored lines, rustication. Some evidence for pottery production was noted in the identification of wasters (or at least 'seconds') with sporled surfaces, and overtired vessels. These include an R2 bowl (Figure 43 , No. 69) and a G2 jar (Figure 43, No. 68), and overtired fl agon (and undiagnostic) sherds in 04.

Brooches by D. Mackreth A total of four complete brooches and one additional fragment were recovered from this site. The complete examples were all recovered from the fill of Pit 22. The assemblage comprises three Colchester Derivatives (Figure 31, Nos 4, 8, 10), one unclassified type (No. 24), and a fragment of a bent pin with two sheet pieces forming part of a circle (No. 36). The brooches are described in greater detail in Chapter 6. The Colchester Derivative examples are similar to those found on Site 1555 in using a rearward-facing hook as the chief visible means of securing the spring. This is considered to be a specifically lcenian style of the mid first century, being restricted to northern East Anglia and consistently dated to the period c. 40-60/65AD (Mackreth 1996). Samian by B. Dickinson Samian was recovered from most features on Site 1515 and the assemblage spans the period from the mid first to the later second centuries AD. Parts of at least twenty-five vessels were recovered, with South Gaulish Forms 15/17,

49

Site 20199 (Figures 2 and 3) It was anticipated that Romano-British features or artefacts would be encountered during the Watching Brief on this site, as surface finds and cropmarks had been

50

observed, but very few features were recorded, either on the roadline itself or in a borrow pit dug to the south . The features visible in this pit did not form part of the general field system, and were partly of more recent date .

Chapter 6. Finds Summaries All artefacts from Site 1555 and the Watching Brief were subjected to analysis at differing levels of intensity. Specialist contributions were generally completed and submitted by late 1992, and have not been consistently updated to reflect subsequent research . The full texts of these reports form part of the archive report (Archive Report Section 4), but summaries are included here. It is intended that these summaries will allow an assessment of the value of the full text for any particular research.

I. Iron Age Coins by Tony Gregory (Archive Report Section 4.1)

Plate XI Iron Age coins: Icenian gold stater; silver unit of Pattern-Horse type; gold quarter-stater. Scale 1:1

Three coins were discovered by metal detectors during the 1989 excavation: an Icenian gold stater, British Na of the latter part of the first century BC; a silver unit of PatternHorse type, likely to have been produced in the second or third quarter of the first century AD; and a gold quarterstater of previously unrecorded type, equivalent to British Ja (Norfolk Wolf) (Plate X1) . The obverse of this last coin showed (off-centre) a curved crest-like motif with curls at edge; the reverse showed a wolf-like animal (facing right). By ana logy with the full denomination this quarter-stater should probably date to the mid first century BC. Two others were found earlier, some distance to the east, though near enough to be regarded as from the same general site (two gold staters of Gallo-Belgic Type C, unpublished but entered in the Index of Celtic coins at the Oxford Institute of Archaeology, datable to the early part of the first century AD). Although the sample is small it is interesting: four of the coins are early, pre-dating the issue of silver, an unusual circumstance which is normally reversed. It is difficult to escape thf" ~onclusion that coins were in use in this area in the first century BC.

Issue

Ila lib

Ill IV V VI VII a V lib VIII IX a IXb X XI Total

11. Roman and Later Coins by John Davies (Archive Report Section 4.2) A total of seventy-eight coins was recovered from the Snettisham Bypass excavation. Fifty-six Roman issues comprise the largest component. There are seventeen post- Roman types and two items are completely illegible (Table 1). All of the coins recovered are very worn and their full identification has been impaired in a number of cases where no surface detail remained. However, it has been possible to identify the majority of issues from their distinctive shapes and sizes, in combination with other diagnostic features such as portrait si lhouettes and isolated lettering.

Period

toAD41 41-54 54-68 69-96 96--117 117-138 138-161 161-180 180--192 193-222 222-238 238-259 259-275 275-296

Total

Total Site 1555)

2

(!)

5 7

(4)

(5)

8 5 6 4

(6)

2

(!)

(2)

(!) (3)

I

41

(23)

12

(10)

first-second century third-fourth century Total Ru111an

3

(2)

:'io

(35)

Iron Age thirteenth-fifteenth century seventeenth century illegible

3 5 12 2

(3)

Coin total

78

(44)

(2)

(3)

(!)

Table 1 Chronological breakdown of the coins from Snettisham Bypass, (Site 1555 totals in brackets)

Roman Issues The bulk of the Roman coins are early in date, with forty-nine of the fifty-six belonging to the first and second centuries, and before. It is unusual for a collection of site finds predominantly to comprise the earlier coins of the

51

Republic Claudius (irreg.)

Denarii (1) 2

Sestertii

Dueondii

2

(2)

There is a gap in the coin list after the late third century until the post-Roman issues begin in the thirteenth century, with a cut halfpenny of King John and a slightly later silver ' long cross' penny of Henry Ill. Other types of interest include three medieval jetons. The earliest of these was produced in the Low Countries in the fourteenth century; the other two are Nurembergjetons, which are commonly found in Britain and date from the fifteenth and sixteenth ce nturies. The remaining post-Roman coins are all halfpennies and farthings dating from Charles II onwards.

( 1)

Nero Vespasian Domitian (1)

Trajan Hadrian

2 2

(1) (1)

(1) (1)

5

(2)

I

3

(3)

2 3 2

(1)

Antoninus Pius

2

Marcus Aurelius/ Luci us Verus

2

(I)

(1)

Illegible first-second

4

(3)

(1)

6

(7)

24

Total

Post-Roman Coinage

Asses

4

(2)

13

(1)

(7)

7

( 1)

(1)

(6)

Ill. Copper Alloy Brooches by D. Mackreth (Archive Report Section 4.3)

(14)

Table 2 Breakdown of first and second-century coin denominations (Site 1555 totals in brackets) Augustan system and these Snettisham coins are numerous enough to allow some numerical analysis. The denominations of the Snettisham first and second century issues are recorded in Table 2. Half of the forty-eight recognisable types are asses, with sestertii accounting for approximately one quarter and dupondii much less common. The comparative rarity of denarii, with just four examples before AD200, is not surprising, as silver coins are always much rarer among site finds. The Snettisham Roman coin list opens with two Republican denarii and two irregular Claudian asses. These particular types comprised much of the coinage used during the immediate post-conquest period and, as such, their presence can often be associated with the site of a military camp, fort or early town. They are found in quantity at the coloniae. Coin supply to Britain was not initially a regular process and following the initial input of Claudian aes during the conquest period, the years between AD52 and the end of Domitian 's reign (AD96) saw only sporadic injection of aes denominations into the province (Walker 1988). There were just four but massive injections of aes during these years, three of which correspond to recognised periods of activity at the mint of Lugdunum. All of the remaining closely datable coins of the first century from Snettisham fall into the first of these periods, between 64-67, under Nero, and all belong to the mint of Lugdunum. After AD96 these was a change in the supply of coin to Britain, with the mint of Rome producing a regular supply of aes. Coin then entered Britain on a more regular basis and there was a heavier emphasis on the production of sestertii while asses became less numerous. These changes are detectable among the Snettisham coins, among which Rome is the only mint represented from the reign of Trajan (from AD98) until the arrival of the antoniniani of the Gallic Empire, in the late third century. Table 2 also clearly shows the change of emphasis towards the production and supply of sestertii from the reign of Trajan onwards. There is a striking absence of late Roman issues which generally dominate on Roman sites in Britain. The main chronological group from Snettisham ends at the start of the third century with a single denarius of Caracalla, struck in AD207. Subsequent issues include just three antoniniani, of AD268-84, and no coin can definitely be assigned to the fourth century.

A total of forty-one complete or fragmentary brooches were studied by D.F. Mackreth; the full descriptive catalogue is included in the archive report. The majority of the brooches were recovered by metal detector during the initial topsoil stripping and were unstratified: the stratified groups have been identified in the main body of the text. All brooches were made from alloys of copper. The majority (21) were classified as Colchester Derivatives, though one 'true' Colchester type was also found (archive catalogue No. 1). This latter brooch has a distinctive design and would appear to have been manufactured somewhere in Norfolk or the southern Fens; although no examples of this design have been dated, the style may indicate the date is late in Colchester'sjloruit: possibly c. AD25-50/55 . The Colchester Derivatives are further classified according to the means originally used to secure the spring and pin to the brooch. Ten used a rearward-facing hook, which is a style believed to be typically Icenian on the basis of its dating (c. AD40-60/65) and northern East Anglian distribution. (Mackreth 1996). Two are of the 'Harlow ' type which are seen as being the successor of the Colchester type in southern East Anglia, and have ajloruit of c. AD40-75/80; and one damaged brooch may have been of the ' Polden Hill' type. Three hybrids of these fixing-styles were identified: hybrids appear to have been very uncommon and the three from the excavation at Snettisham cannot be easily sourced or dated, though it is unlikely that they date to after c. AD80. A further four brooches had hinged pins, three also having moulded decoration - a combination which is very unusual and apparently restricted to East Anglia and the eastern Midlands. The distribution and style suggest that this group too is probably earlier than AD60/65, and certainly no later than c. AD75 . The other identified brooches consisted of four early post-Conquest Hod Hill types, two penannulars, two late La Tene types, and a Headstud type. All these types appear to have gone out of circulation before the end of the first century AD; no brooches were found which could be dated with certainty to the second century or later. House 2 and Enclosure (Phase 2) (Figure 30)

Colchester Derivatives 14. The spring is held by three rearward-facing hooks, one in the middle and one at the end of each wing. The hook has a buried ridge down the middle. The wings and bow are plain. (4 1 (First phase of enclosure ditch fill). SF 43)

52

18

, 'I

i

I

I' '~

;:

19

I

I

I

I

\Q :11

r

:j

J 31

li 25

• 26

Se m

0

Figure 30 Collection of brooches from House 2 Enclosure ditch . Scale 1: 1

53

8

\\ 10

\\

0

5cms

Figure 31 Brooch assemblage, Site 1515. Scale 1:1 usual two-part foot-knob. (187 (First phase of enclosure ditch fill). SF47) 32. The upper bow has a flaring moulding above three thinner ones each separated by straight secti om with chamfered sides. The lower bow has a rounded fro nt. The foot-knob is missing. The tinning or si lvering was applied differentially leav ing a section with the three thin mouldings plain. (4 1 (First phase ofenclosure ditch fill). SF31)

18. The spring is held to the body of the brooch by an axis bar passing through the coils of the spring and the lower of two holes in a plate behind the head of the bow; the upper hole holds the chord. Each wing is plain. The plate behind the head of the bow is carried over that to form a skeuomorph of the hook on the Colchester type. The bow has a fl at back and squared sides. The front has a concave face on each si de and line of rocker-arm ornament down the flat centre. The catch plate is solid. (187 (First phase enclosure ditch fill).) A virtuall y identical brooch, but with a pierced catchplate was recovered during topsoil removal in the southern part of the excavati on area (No. 17, not illustrated). 19. There may be seams across the back of the wings which would show that the ax is bar had been inserted after casting. Each wing has a single wavy line ridge next to the bow formed by punching. Between that and the end of the wing are two pairs of opposed wavy lines. The bow has a flat back, squared corners and a step down each side of the front. The centre of that is rounded and has a pair of opposed wavy ridges sunk down the middle. (22 (Second phase of ditch fill). SF35) 20. There is a seam across the back of the wings showing that the axis bar was inserted after casting. Each wing has three beaJed ridges alternati ng wi th pairs of ridges joined by flutes. The bow has the sa me secti on as 19. above, on ly with a beaded ridge down the middle. (4 1 (First phase of enclosure ditchfill) . SF29)

Headstud 26. The axis bar of the spring passes through a pierced lug behind the head of the bow, the chord is he ld by a forward-facing hook. Each wing has two mouldings, one wide and the other narrow. The c ircular stud has an annul ar groove. Each border below the stud has a groove and there are nine rectangular cell s down the middle. The enamel has completely decayed, but its condi ti on suggests that there had been two alternating colours. At the base of the bow are cross-mouldings separated from the two-part foot-knob by a flute. (22 (Second phase of enclosure ditch fill). SF27) Penannular 33. The ring has a lozenge section and cross-cuts along the top corner. Each terminal was hammered out and then coi led at right angle to the plane of the ring. The pin has a slight hump. (4 1 (First phase of enclosure ditch fill). SF31) Unclassified 25. The head is missing. The plain bow is thin and has a rounded front. The catchplate has five piercings arranged as a crude type of frett ing. Although the head, with its diagnostic characteristics, is mi ss ing the elaborate ly decorated catchplate appears to be a

Hod Hill Types 31. The upper bow has four equally spaced ridges, separated by flutes, the middle ones being beaded. The lower bow has a flute and a moulding at the top and a line of cross-cuts down the middle to the

54

deliberate attempt to copy the style of elaborate fretting found on some Colchesters (e.g. Stead and Rigby 1989, fig .125,9,11). This would imply that there were still examples to be seen in use and wou ld make the date of manufacture to before 55/60 at the latest. (167 (Fill of house ditch, northeast sector). 167)

cooling surface. The tap slag from Site 1555 was typical and the majority was readily distinguished from the smithing slags. A total of 27.6kg was recorded. The slag ranged in size from fragments< 50 mm maximum to larger pieces < 100mm maximum. A few very large lumps weighing in excess of 200g were recovered. Examples of 'slag feeders' (runs of slag that solidified in the furnace tap hole) were recovered from the Phase 5 backfill of Well I.

Rubbish pit group (Phase 2)

Colchester type n.ill. (No. I ) Only the stubs of the spring and hook survive. The remaining wing is very short and plain. The bow is wide and has a slight taper towards the foot, now missing. Down the centre of the bow are two ridges distorted by punching to form opposed wavy lines. (357 (Fill of Pit420). SF 54)

Furnace Slag This smelting slag is characterised by its viscous appearance, and the presence of large charcoal impressions (approximately 25mm in square section and at least 30mm long). There were only a few examples of this slag type from cleaning to the north of Well 1, and the material has not been recorded separately in the archive catalogue (Archive Report Section 4.4).

Colchester Derivatives n.ill. (No. 5) Very corroded, the front face of the wings is hidden. The bow has a flat back, a step down each side and a beaded ridge down the centre of the swelled fro nt. (357 (Fill of Pit420) . SF 53) n.ill. (No.15) The spring is held by three rearward-facing hooks. The hooks are carried dow n to form ridges, the central one being short, the others running to the bottom of the wings and having another ridge on the inner side. The bow has a ridge between it and each wing. the bow itself is plain and the lower part is missing. (357 (Fill of Pit420). SF 55)

Smithing Slag This classification was used for all slag not readily identifiable as smelting by-product; in excess of 2.4kg of this type of slag was noted .

Site 1515 (Figure 31)

Hearth Bottoms A piano-convex accumulation of iron silicate forming in the smelting hearth . Only one example was present, coming from Pit 738, part of the Phase 3 iron processing concentration discovered during the Watching Brief at the north end of Site 1555.

Colchester Derivatives 4. Each wing has two sunken ridges, separated by a flute , at the end. The rest of the brooch is pl ain. (23, Fill of Pit 22) 8. The surviving parts of the wings are plain. The bow has a flat back and a rounded front down which runs a sunken beaded ridge. (23, fill of Pit 22) 10. The complete wing has two pairs of sunken ridges divideu uy a flute between three sets of two opposed wavy line formed by distorting a pair of buried ridges by punching. The hook is broad with two buried ridges forming a pair of opposed Cs on top. At the foot of the hook is a simplified set of mouldings copying those on the wings. The bow has a fl at back, chamfered rear corners, a step down each side and two sunken wavy ridges, formed like those on the wings, down the rounded front. (23, fill of Pit 22)

Non-Diagnostic Slags and Residues Hearth Lining The clay lining of an industrial hearth/furnace/kiln with a vitrified or slag-atta~ked face. It is often impossible to distinguish between furnace and hearth lining. A total of 3.7 kg of this material was recovered from the 1989 fieldwork, with a sub-rectangular furnace lining fragment being found in the third/fourth-century backfill of We Ill. This has been described more fully above in Chapter 4.

Unclassified 24. Is missing its head ; the bow has a lozenge section with two pairs of ridges down the middle. The catchplate has two large piercings which are fretted next to the elaborately-shaped dividing bar which is pierced by a small circle. The start of the return has a vertical groove, and the return itself has a relief ' lip ' at the top and, below that, a chevron made up of pairs of grooves. Although the head is missing the elaborately decorated catchplate should be no later than c. 75AD. (23, fill of Pit 22)

Cinder High silica-content slag, which can form either at the reaction zone between the smithing slag and the hearth lining, or by high temperature reaction between silica and ferruginous material. Only 67g of this class of slag was recovered .

IV. Metalworking Debris Slag by Gerry McDonnell (Archive Report Section 4.4)

Other Residues This cl ass comprised possible ore fragments, ferruginous concretion (probably naturally formed), and fragments of fired clay. 1.0kg of this material was recorded.

Slag Classification The slag recovered from the excavation of Site 1555 and from Watching Brief sites amounted to 34.8kg in total. The material was visually examined and classified solely on the basis of its morphology into two broad groups: diagnostic slags (i.e. those which can be attributed to a particular industrial process) and non-diagnostic, which could have been produced by a variety of processes.

Distribution of the Slag The s la g recovered during detailed excavation, fieldwalking, and in the Watching Brief has been catalogued in the Archive Report (Section 4.4). Slag was recorded in a total of seventy-five contexts along the whole length of the excavation; on ly a few contained a weight of slag greater than 1kg (a figure taken to represent potentially significant deposits). The greatest quantities of tap slag were recovered from the unstratified layers, with appreciable stratified quantities coming only from the backfill of Well 1. The di stribution of the hearth/furnace lining reflected that of the smelting slag and it is therefore

Diagnostic Ferrous Slags and Residues Smelting Tap Slag Tap slag is one of the most characteristic forms and is distinguished by the rope-like morphology of the upper

55

probable that the lining is furnace lining, which would indicate that the smelting area was close to the part of Site 1555 excavated in 1989. The quantities of smithing slag, cinder and other material are background levels only.

by a relatively fine-textured and clean fabric . Frequently the outsides of moulds are oxidised through being in contact with air whilst at elevated te mperatures (i.e. during pre-heating and casting). However a number of moulds from the excavation had thinner sections or were exposed to the reducing atmosphere for longer so that their fabrics were reduced throughout.

Discussion The distribution of the slag in excavated features showed no concentrations pointing to areas of smelting activity. The occurrence of large deposits in the plough soil on Site 1555 and in large features suggests that the slag has spread from an ironworking area close to the excavation . The slag is typical tap smelting slag, which is common in the Roman period, although the presence of a piece of probable furnace lining which was only slightly curved does suggest the presence of furnace types other than the most common shaft furnaces of the ' Ashwicken' type (Tylecote and Owles 1960).

Identifiable Forms (Figure 32) Although fragments are generally small, some can be identified as parts of piece moulds.

1.

Conclusions The evidence recovered from excavation and field walking indicates that there was Roman iron smelting close to the line of the present day Snettisham bypass. It is possible that the activity could be earlier, and the slag occurring in the Roman contexts is residual. The scale of the activity cannot be determined, i.e. there is insufficient evidence to suggest whether there were just a few smelting operations to sati sfy an immedi ate need or there was a minor industrial activity. The slag is typical iron smelting tap slag with some furnace slag present.

2.

Moulds and Crucibles by C. Mortimer (Archive Report Section 4.5) Approximately eighty small sherds recovered from the Phase 1 industrial waste pit 730 were examined . The material consisted of various forms in various degrees of oxidation, and was categorised as crucible or mould fragments on the basis of form and degree of oxidation. Many of the pieces whose surfaces are predominantly reduced-fired are thought to be crucible fragments . On the whole they are too small for their origi nal forms to be reconstructed, or their diameters estimated. The crosssections are rather irregul ar so they were probably hand built. They may have been dish- or bowl-shaped with rims pinched out to form pouring lips. Most crucibles found in the 1989 excavations are thick (up to 12mm) and their fabrics have a somewhat coarser texture than the moulds, sometimes with large inclusions. The forms discovered at this site are consistent with late Iron Age forms (Spratling et al. 1980, fig. 2). Vitrification is evident at the edge of several pieces, either on the inside or on both the inside and outside of the fragments. This may suggest that heating was from above. A selection of these areas was tested using qualitative non-destructive x-ray fluorescence (XRF), as were a smaller number of metal-rich deposits and some inside surfaces with unusual colouring. A representative group of results is recorded below in the Metalworking Debri s Summary, and in the Archive Report (Section 4.5). The XRF results suggest that copper alloys containing zinc, lead and tin were being melted at this site, though the exact alloy(s) involved cannot be determined. Mould fragments are identified by the presence of object impressions (the area of the impression normally being strongly reduced by contact with the hot metal), and

3.

Mould fragme nt apparently for a ring, 60-70mm in di ameter, with rounded elements projecting from it. It is suggested that the mould would be suitable for casting terrets. Pieces from the upper and lower valves of this mould can be di stingui shed by the nature of the outer rim of the mould ; these areas may be either concave or convex and would have fitted together to make a smooth join. Whe n fitted together th e mo uld a ppears to have been doughnut- shaped externall y (i.e. annul ar with a very small central hole). Several pieces of a tapering cylinder with minimum internal diameter of slig htl y less than 20rnrn, flaring out to a diameter of at least 40mm. Although this is the type of profile expected for a sprue c up, the di ameter is too large. It is difficult to know which end was uppermost in casting; at the wider end the smoothed outside surfaces of the mould are consiste ntly broken off, a feature which may relate to some sort of luting which joined thi s part of the mould to another part, causing the surface to be pulled away when the mould was taken apart. At the other end of several of the pieces there is a slight ridge whi ch would suggest a mould form with a broad rounded bulge, about 5rnrn wide. This ridge is also vi sible on other pieces in the assemblage. If these pieces are indeed assoc iated, the form ev idently continued into a curving shape, with a much sharper curvature than the flared piece discussed above. A possible reconstruction has been suggested (Figure 32.2). With thi s size of object, it is li kely that the casting would have been hollow and thu s required an inner ceramic core. Some pieces with this detail are oxidi sed and others reduced on the su rface, suggesting that they are not all from the same casting operati on, and therefore that thi s form may have been made more than once at thi s site. Mould fragment with a flat base and an almost sq uare corner. Inside the base there seems to be a sli ght upturn. This piece could be from a bar ingot mou ld, in which case this upturn may be the lower part of an outside wall. An ingot mould is quite likely, as thi s piece appears to be from an open mould.

Subseq uent to the metalworking debri s analys is four fragme nts of possibl y three additional moulds were identified during study of the fired clay from the 'cleaning' context in the same area. 4. Mould fragment consisting of a concave terminal banded by a ridge, with an apparentl y curving shank extending beyond the edge of the fragment. The edge surface of the mould, if any, has been lost, as has the continuati on to the shank. 5, 6. Fragments of mould for an essenti ally fl at object, with a smooth everted surface. The edges of the mould appear to have been broken off, leav ing no traces on the fragments in Figure 32.6. Figure 32.5 shows part of an object with a tapering longitudinal ridge; the two fragments in Figure 32.6 appear to be two related parts, possibly of the sa me object as Figure 32.5 on the basis of their shape.

A selection of vitrified mould and crucible fragments, and a smaller number of metal-rich and anomalously coloured fragments were subj ected to qualitative nondestructive x-ray fluorescence to determine the constituent metals (Table 3). The vitreous areas gave little indication of non-ferrous metal types, but the metallic areas and inside surfaces were more illuminating. The XRF results suggest that copper alloys containing zinc and lead were being melted at this site. (A more complete di scussion of the results is provided in the Archive Report Section 4.5). In conclusion we can say that copper alloys containing zinc, tin and lead were melted and cast on the site, producing a number of artefact types. It is likely that the crucible and mould fragments came fro m the same metalworking system.

56

2

IOcm

0

3

4

t

5

6

Figure 32 Copper alloy mould fragments from pit 730. Scale 1:1

XRF Analysis of surfaces: Crucibles Context Area 665 Metallic lump 730 Large frag, metal blobs Slag, red and green Pouring lip- orange Slag Inside a pouring_! ip

XRF Analysis of surfaces: Moulds Context Area 730 ?terret (inside) terret ?fitting corner ' ingot' piece

Results Pb Cu Zn Cu Zn Pb Sn Cu Sn Pb Cu Pb Zn Pb Cu Sn ZnPb

Results ZnPb Zn Pb (Cu) Pb Zn (Cu) Cu Pb

Table 3 XRF analysis of moulds and crucibles

Note: Elements are li sted approximately in order of peak heights in XRF spectra. Those in brackets are present in very low quantities.

57

V. Glass

I

by J. Price and S. Cottam (Archive Report Section 4.6) The excavations at Snettisham produced thirty-nine fragments from seven vessels. Thirty-three of these were colourless, and six blue/green. Heavy iridescence was present on the group of fragments making up a cylindrical bottle (No. 2 in Archive Report and Figure 33), but the remaining fragments , particularly those of blue/green glass, showed little weathering. The seven vessels represented comprised a colourless cup or beaker (No. 1), a cylindrical bottle (No. 2), two possible flask or unguent bottles (Nos 3, 4), one cylindrical and one prismatic blue/green bottles (Nos 5, 6), and an unidentifible sherd (No. 7). The cup and clear bottle both appear to be of types current in the second century, though the bottle is similar to types manufactured from the second to the fourth centuries.

/

I

1

JK I

'

r

Illustrated Glass Vessels (Figure 33) 1. Thirteen body fragments, many joining, probably from a beaker. Colourless. Straight side taperi ng in to constriction, then expanding out. Part of two indents on si de below constricti on. Horizontal marvered trail at constriction. Small bubbles. Eound in the late third-century back fill (634) of Well I. Fragments from the rim and base of a similar indented cup, with horizontal trail below the rim, were found in a second-century context in the drain deposit of the Commandant 's House at tfousesteads (Charles worth 1971, 34, no.

L_-----iL~-~J 2

6).

2.

Twenty fragments, shou lder, body and base, cylindrical bottle. Pale green/colourless. Slight constri ction on body below shoulder. Concave base. Two hori zontal abraded bands on upper body. At least two further bands on body. Vertical scratch marks. Ring of wear on base edge. Patchy, flaking irridescence. Some fragme nts pitted. Recovered from a complex of rubbi sh pits (128) close to Well I. Colourless vessels with cylindrical bodies occur at various times from the second to the fourth centuries. For instance, simil ar though smaller vessels with either two 'dolphin' handles or no handles, or no neck and a hole-mouth with a tubular edge, are known from York (Harden 1962, 140, fig. 90 H.G.I46.1-4 and 141 fig . 89 H.G.l82). The shape of the shoulder fragment indi cates that the Snetti sham bottle had a neck, and it is likel y that these fragments come from a large bottle with one angular handle, although no evidence for a handle now survives.

3

Figure 33 Glass vessels, Site 1555. Scale 1:1

would have had an out-turned rim with a sheared edge, a long narrow neck with a tooled constricti on between neck and body, a gently expanding body and a convex base. Production of thi s form seems to have ended in the early Flavian period. 5, 6. Both these fragment s come from blue/green bottles which were produced for the containment of liquid and semi- liquid substances and were in common use throug hout the Roman world during the first and second centuries. No. 5 comes from a cy lindrical bottle (!sings 1957, Form 51) found in a Phase 3 ditch fill , and No. 6 from a pri smat ic bottle from Phase 2.

Two of the be s t examp les of thi s type come from Hauxt on, Cambridgeshire (Harden 1958, 13 fig. 6) and Corbridge (Charlesworth 1959, 54 fig. 10 no. 1). Fragments from similar vessels have been found at Caerwent (Boon 1974, 112 no. 3a-